A new method to calculate
the geodiversity index of
volcanic hotspot islands: The
Hawaiian archipelago
Author: Jorinde Guldenaar
10799753
01-07-2017, Amsterdam
Supervisor: Dhr. Dr. A. C. (Harry) Seijmonsbergen
Bsc Earth Science Thesis
Institute for Interdisciplinairy Studies
The figure at the cover visualizes the geodiversity of the Hawaiian archipelago for a 500x500m grid and a particular geodiversity index that was examined in this research. Low levels of geodiversity are characterized by green colours, while high geodiversity is indicated by red colours.
Abstract
The term geodiversity is used to describe the abiotic nature in a certain area. This research aims to calculate a geodiversity index (GDI) for volcanic hotspot islands and compare the geodiversity between islands of the Hawaiian archipelago related to their stage in island development. The life span of an island is also referred to as the Wilson cycle. This research will tackle nine different geodiversity scenarios for the Hawaiian Islands including three different grid sizes (500m, 750m, 1km) and three different GDI indices, through the use of Geographic Information Systems (GIS). Geology, pedology, hydrology, elevation and slope maps are overlaying by a grid in order to obtain the diversity in each grid cell through the calculation of zonal statistics. These sub-indices are used as the input to the GDI indices. After analysing the results, the 500m grid in combination with the formula GDI1 was suggested to be the most appropriate for the Hawaiian archipelago.
The geodiversity index reveals an increasing pattern of high geodiversity from the younger islands, located in early stages, to the older Hawaiian Islands, located in later stages of the Wilson cycle. The geodiversity is increasing throughout the Wilson cycle. The older the islands, the higher the
geodiversity. The factor time plays an important role in the understanding of this pattern, since there is a co-evolution of topography and hydrology over the archipelago.
However, conclusions of geodiversity of islands in very early stages of the Wilson cycle cannot be done yet. It is therefore suggested that the method presented in this research has to be applied on other hotspot island chains where it is possible to calculate geodiversity in early stages of island development. Besides, further research could be carried out to geodiversity at different tectonic settings, comparing geodiversity and biodiversity hotspot on island chains or connecting the geodiversity of islands with the seafloor geodiversity in order to explore geodiversity on a more global scale.
Contents
1. Introduction
5
1.1
Geodiversity: relevance and innovative aspects
5
1.2
Objective, research questions and hypotheses
6
2. Theoretical framework
72.1
Study area: The Hawaiian archipelago and the Wilson cycle
73. Methods
8
3.1
Research workflow
8
3.2
Data collection
8
3.3
Data pre-processing
8
3.4
Development of the Geodiversity Index
9
3.5
Calculating the sub-indices
9
3.6
Calculation of the Geodiversity Index
10
3.7
The analyses in ArcGIS
11
4. Results
12
4.1
The geodiversity maps
12
4.2
The analyses
13
5. Discussion
17
5.1
Methodology
17
5.2
Interpretation of the results
17
5.3
Further research
18
6. Conclusion
19Acknowledgements
20References
21Appendices
23Appendix A: The physiography of the Hawaiian Islands
23
Appendix B: Input maps
24
Appendix C: Sub-indices 500m grid
29
Appendix D: Geodiversity maps
30
Chapter 1
Introduction
1.1
Geodiversity: relevance and innovative aspects
Geodiversity is a relatively young term in the Earth Science. Since the 1990s the term geodiversity is used to describe the variety of abiotic nature in a certain area (Gray, 2004; Serrano & Ruiz-Flaño, 2007). Geodiversity was defined by Gray (2004) as ‘the natural range (diversity) of geological (rocks, minerals, fossils), geomorphological (land form, processes) and soil features. It includes their assemblages, relationships, properties, interpretations and systems’. It is equivalent to the term biodiversity, which describes the variety in biotic sphere.
In the last decade various research was done on measuring geodiversity. Serrano et al. (2009) developed a geodiversity index (GDI) of the rural landscape of Spain. Many attempts have been done to improve the GDI (Hjort & Luoto, 2010; Perreira et al., 2013). Chambers (2014) developed a geodiversity index for the hotspot island Tenerife. The formula differentiates from other indices because it takes into account slope coefficients to characterize the shape of volcanic islands.
Proposed is an innovative study on the calculation of a geodiversity index and comparison of the geodiversity between volcanic islands in the Hawaiian archipelago related to their stage in island development. This research will tackle nine different geodiversity scenarios for the Hawaiian Islands including three different grid sizes and geodiversity indices. The Hawaiian archipelago is considered to be an excellent study area for the purpose of this research, because the islands created by a stationary hotspot, are showing different stages of oceanic island ontogeny (Mergurian & Okulewicz, 2007; Wilson, 1963).
Measuring geodiversity is fundamental for the purpose of geoconservation, which is ‘the action taken with the intent of conserving, understanding and valuing geological and geomorphological features, processes, sites and specimens’ (Burek & Prosser, 2008). In general most people may assume that the abiotic environment of rocks and mountains is very static and that there is no possibility that it will be disturbed. On the contrary, the opposite takes place. Whereas biodiversity is mostly able to recover after disturbance, geodiversity is not. The disturbance of geodiversity will cause irreversible damage (Gray, 2004). Geodiversity has to be seen as a valuable natural heritage and deserves to be conserved for future generations (Burek & Prosser, 2008).
A major aim in recent island studies has been to discover the species patterns through time in the life span of volcanic islands. Whittaker, Triantis & Ladle (2008) did research to a model of oceanic island biogeography that may explain biodiversity patterns by taking also the island ontogeny into account. The development of a new volcanic island begins with the growth towards a highly elevated cone-shaped geometry above sea level. Maximum topographic complexity is suggested to occur after this period and is followed by degradation of the islands towards an almost flat eroded surface. This life cycle of an oceanic island is also referred to as the Wilson cycle (Wilson, 1963).
Examination of variation in geodiversity of islands in different stages of island development may be an important key to the understanding of diversity patterns on archipelagos. Kozlowski (2004) stated that the condition of the environment of both the lithosphere and the earth surface are intrinsically linked with the creation, development and destruction of life. If a geodiversity index for the Hawaiian Islands is successfully developed, the method could be applied on other archipelagos and be compared with biodiversity patterns.
The objective, research questions and hypothesis of this study are described in the following paragraph. Furthermore, the theoretical framework will be described in chapter 2. In chapter 3 the methods will be presented. The results will be presented in chapter 4 and discussed in chapter 5. Lastly, a conclusion will be drawn on the geodiversity of volcanic hotspot islands in chapter 6.
1.2
Objective, research questions and hypothesis
The research aims to calculate and compare the geodiversity of the Hawaiian Islands. The following main question and sub-questions are aspired to be answered:
What is the relationship between the geodiversity of the Hawaiian islands and their stage in oceanic island development?
Q1. What grid size should be used for the calculation of the geodiversity of the Hawaiian islands? Q2. In what stage of oceanic island development, i.e. the Wilson cycle, can the eight major Hawaiian Islands be placed?
Q3. What is the most appropriate geodiversity index to calculate the geodiversity of the Hawaiian Islands?
Q4. What geodiversity patterns can be observed between the islands of the Hawaiian archipelago? Hypothesis:
The geodiversity will first increase in north-western direction on the younger islands. The highest geodiversity can be found on the island with maximum topographic complexity, expected to be Maui’i, northwest of the island with the highest area and elevation (Hawai’i). This will be followed by a decrease in north-western direction over the older islands.
Chapter 2
Theoretical framework
2.1 Study area: The Hawaiian archipelago and the Wilson cycle
In this paragraph a brief introduction to the physiography of the Hawaiian archipelago will be presented. This is important to understand as geodiversity is thought to relate to the different tectonic stages in island development.
In the case of the Hawaiian archipelago the Pacific Ocean plate is drifting in north-western direction over a hotspot that is creating the islands (see Appendix A, geological setting Hawaii). The hotspot is a relatively stationary hot plume in the earth’s mantle. The islands diminish in size due to subsidence and erosion as they are further away from the hotspot that created them (Mergurian & Okulewicz, 2007).
Wilson (1963) suggested that ‘Volcanoes of the Hawaiian chain had similar, rather than identical, histories and each volcanic island in turn went through a similar cycle of volcanism and erosion, one after the other’. Each island of the archipelago belongs thus to a particular stage in the Wilson cycle and may be related to geodiversity patterns. The life cycle of an oceanic island begins with growth towards a highly elevated island of maximum growth after which a period of degradation, erosion and subsidence will result in an almost flat eroded surface (see Appendix A, the Wilson cycle).
During a Wilson cycle, maximum topographic complexity on an oceanic island will occur after periods of highest elevation and largest area (fig. 1) (Whittaker et al., 2008). In areas with a high complex topography it is suggested that there is a wide variation in slopes, affecting the soil formation, a wide variation of elevation causing large variety in geomorphology. Areas with high complex topography are therefore expected to be related with high and very high geodiversity.
However, geodiversity includes expert-based information such as geological, geomorphology and soil features (Gray, 2004) which are dependent on the factor time. Soils become more complex towards the older, more stable islands. Geomorphology is also developing through time. The younger islands consists of a smooth surface, because they are located above an active hotspot, which is causing lava flows filling up the river valleys. A decline of active volcanic processes in combination with an increase of degradation towards the older islands may also be the reason for roughness in the landscape (Jefferson, 2014).
Figure 1: Ideal and simplified relationship between age (x-as) and area (dotted line), elevation (dashed line) and topographic complexity (solid line) of an oceanic hotspot island (Whittaker et al., 2008).
Chapter 3
Methods
3.1 Research workflow
A simple overview of the workflow required for the mapping process is presented in figure 2. The work flow consists of 7 general steps, that are further discussed in separate paragraphs. Step 1 to 6 are discussed in this chapter. The visualization of the GDI as well as the results of the analysis are presented in chapter 4. Step 8 is discussed in chapter 5.
Figure 2: The research workflow
3.2
Data collection
For developing a geodiversity index of the Hawaiian islands digital map datasets had to be collected
(table 1). Geology and pedology maps of the Hawaiian Islands were available (Sherrod, Sinton, Watkins
& Brunt, 2007; USDA, 2015). A geomorphology map of the study area did not exist. Therefore, a 10 m DEM of the Hawaiian Islands was used (Greenberg, 2007) to calculate the elevation and slope diversity. Lastly, the hydrology diversity was computed by using datasets containing rivers and waterbodies (USGS, 2010).
Table 1: Overview of the metadata and its sources
3.3 Data pre-processing
The data presented in table 1 has been pre-processed by overlaying the data with three different grid sizes (1x1 km, 750x750m and 500x500m) that will cover the whole island using the fishnet tool. Next, the grids were clipped to the boundary of the islands. A 1km grid can be used with the aim to reduce computational time. However, more detailed results can be achieved by using smaller grid sizes since fine input sizes will be used. Therefore the 1 km grid is halved to experiment also with a 500m grid and the 750m grid was used to examine the result somewhere in between.
Consequently, incomplete grid cells at the border were removed, because the GDI can only be calculated for cells of the same size. For the hydrology the two different datasets containing waterbodies and rivers were combined using the union tool. The geology, pedology and hydrology maps were
Sub-index Contains Data type Coordination system Scale/ cell size Nr. of different features Publication date Source
Geology Geological units Polygon NAD83_HARN_UTM_zone_4N 1:250.000 9862 2007 USGS Pedology Soil units Polygon GCS_WGS_1984 1:250.000 378 2015 USDA DEM Elevation of the
surface
Raster NAD83_HARN_UTM_zone_4N 10 x 10 m 0-4200 m 2007 NOS & NCOSS Hydrology (streams) River streams (perennial and non-perennial) Polygon GCS_WGS_1984 1:24.000 1200 2010 USGS Hydrology (waterbodies) Lakes, ponds, reservoirs, waterholes Polygon GCS_WGS_1984 1:24.000 3194 2010 USGS
converted to rasters of 10 meter cell size using the convert feature to raster tool. A cell size of 10 meter is chosen based on the cell size of the DEM.
The hydrology was further pre-processed, because perennial and non-perennial rivers had to be distinguished. A separate raster containing waterbodies and perennial rivers was created. In another raster the non-perennial rivers were stored in order to give them different weightings in the following steps. The reason for this is that perennial rivers are suggested to have a larger influence on the geodiversity, because they have water flow all year round. On the contrary, non-perennial rivers only flow half a year (NOAA, 2016).
3.4 Development of the Geodiversity Index
A Geodiversity Index (GDI) was created for the Hawaiian Islands. The software program ESRI ArcMap 10.1 was used for data analysis. Nine scenarios, including three different grid sizes and three different formulas, were carried out in order to determine the most appropriate formula for calculating the geodiversity index. The following GDI formulas were tested:
Formulas 1.1, 1.2 and 1.3 consist of small adjustments to examine the effect of slope and elevation diversity on the resulting geodiversity index, since the geomorphology is lacking. According to Ascione, Cinque, Miccadei, Villani & Berti (2008), the standard deviation of elevation is a good measure for topographic roughness. In contrast Ruszkiczay-Rüdiger, Fodor, Horváth & Telbisz (2009) suggest that
the range of elevation in a certain area is the basic topographic parameter.The first and second formula
take into account the standard deviation or range of both land surface parameters separately. The third formula includes a combination of both range and standard deviation. The slope and elevation diversity are multiplied by the value 0.5, because the calculations of the range and standard deviation of slope and elevation diversity then have an equal weight in comparison to the other indices.
3.5 Calculating the sub-indices
The following steps were executed three times for the three different grid sizes. The geology, pedology and hydrology raster maps (see Appendix B) together with the grid were used as inputs to the zonal statistics tool in order to calculate the amount of different legend items found in each quadrant. The geology and pedology were divided into five classes of equal intervals with the reclassify tool. The two hydrology rasters were reclassified to give the perennial rivers and waterbodies value 2 and the non-perennial rivers value 1. Water is thought to contribute more to processes such as erosion and mass movement, which create the variety in landforms, elevation and resurface roughness. That is why the perennial rivers and waterbodies, were given a value two times larger than the non-perennial as the rain season in the Hawaiian archipelago is only half a year (NOAA, 2016). Consequently the two rasters were added up using the raster calculator.
This resulted in three sub-indices, Gdi, Pdi and Hdi, which show the variety for each grid cell. The elevation diversity was calculated from the DEM directly by selecting standard deviation (Esd) and range (Er) in the zonal statistics tool. For calculating the slope diversity (Ssd/Sr) the slope tool was first used to create a slope angle map from the DEM. The input for the zonal statistics tool was the standard deviation or range. The reclassify tool was used to make five classes of equal intervals for the Ssd, Sr, Esd and Er. The range is the difference between the minimum and maximum value in each grid cell. A high value represents a large difference between the steepest and shallowest slopes within a quadrant or a large difference between highest and lowest elevation. The standard deviation expresses the variation within a quadrant which indicates, in case of a high value, that there is a large variation in slopes or elevations. High values of these parameters thus point out that there is high importance for geodiversity. The
minimum and maximum scores of the varieties (Gdi, Pdi, Hdi), standard deviations (Ssd, Esd) and ranges (Sr, Er) before reclassifying are presented in table 2.
Table 2: Showing the minimum and maximum scores of the parameters used for the GDI formulas, before reclassifying
[min-max]
Pdi Gdi Hdi Er Sr Esd Ssd 500m grid 1-7 1-6 0-3 0-255 0-83 0-245 0-24,5379 750m grid 1-7 1-6 0-3 0-1029 0-83 0-255 0-23,9084 1km grid 1-8 1-7 0-3 0-255 0-84 0-255 0-24,4597
3.6 Calculation of the Geodiversity Index
The data of the five sub-indices were used as the input to formula 1.1, 1.2 and 1.3 (raster calculator tool) to obtain a geodiversity value for each grid cell. The classification method for the GDI is the natural breaks (Jenks) method. This method divides each grid cell in a geodiversity class with the slice tool into: very low (1), low (2), medium (3), high (4) and very high (5). The method of classification used by Jenks (1967) serves to minimize the standard deviation of values within a class, while
maximizing the standard deviation from the means of other classes. The Jenks method creates categories that have comparable levels of geodiversity. This result in nine GDIs with five classes representing different geodiversity values (tables 3, 4, 5) A simplistic visualization of the calculation of the GDI is presented in figure 3.
Table 3: Scores used for the classification of the geodiversity, based on the 500m grid GDIs
500m grid GDI value
GDI1 GDI2 GDI3
1 4-5 4-5 4-4,952941176
2 6-7 6-7 4,952941177-6,964705882
3 8-10 8-10 6,964705883-9,505882343
4 11-12 11-12 9,505882344-11,51764706
5 13-18 13-18 11,51764707-17,5
Table 4: Scores used for the classification of the geodiversity, based on the 750m grid GDIs
750 m grid GDI value
GDI1 GDI2 GDI3
1 4 4-5 3 - 3,980392157
2 5-6 6-8 3,980392158 - 5,5
3 7-8 9-11 5,500000001 - 7,509803922
4 9-10 12-14 7,509803923 - 9,519607843
5 11-14 15-21 9,519607844 - 15,5
Table 5: Scores used for the classification of the geodiversity, based on the 1km grid GDIs
1km grid GDI value
GDI1 GDI2 GDI3
1 4-5 4-6 5-6
2 6-7 7-8 7-8
3 8-9 9-11 9-10
4 10-12 12-14 11-12
5 13-17 15-21 13-16
3.7 The analyses in ArcGIS
The following analyses, based on the geodiversity maps of all islands, are made: 1. The effect of the sub-indices on the GDI, 2. Comparison of the standard deviations per island and 3. Comparison of
percentages of all geodiversity classes per island.
The band collection statistics tool was used for calculating correlation matrices, in order to evaluate the effect of each sub-index on the GDI by presenting the degree of linear dependence between the indices. Values vary from 0 to 1. The higher the correlation value, the stronger the positive
correlation between two sub-indices. This means that if values of the sub-index are increasing, the values of the other sub-index are increasing as well.
Features that only contain the boundaries of the three different fishnets are necessary to compute statistics of each island by using the zonal statistics as table tool. Since we are only interested in the boundary of the fishnets, the grid cells inside this boundary had to be removed. This result serves as the input to the zonal statistics as table tool to calculate statistics of the three GDIs over the whole archipelago. For the output the option ‘ALL’ is selected, calculating among others the mean and standard deviations of each island. As a consequence the optimal grid size is selected. This selection was based on the method of Hengl (2006) who used the output with the highest levels of variation as a guideline for selecting the most appropriate cell size.
For calculating percentages of geodiversity classes of each island, the boundary layer of the fishnet was clipped for every island. In the zonal statistics as table tool, the boundary of each island was the input and the output raster was the GDI layer. This resulted in the geodiversity values 1 to 5 in combination with the count of each value. The percentages of all GDI classes was calculated by dividing the count of a certain value by the total number of counts and multiply it by 100.
Chapter 4
Results
4.1 The Geodiversity maps
In this paragraph the results of the geodiversity indices and the statistical analysis are presented. Figure 4 shows that the GDIs differ in degree of variation of geodiversity over the Hawaiian archipelago. An increase of high geodiversity in north-western direction can be detected for all GDIs.
Figure 4: Geodiversity maps of the Hawaiian Islands for three different grid sizes (500m, 750m, 1km) and for three different GDI formulas, classified from very low to very high geodiversity.
GDI 1
GDI 2
GDI 3
500m grid 750m grid 1km grid
The results of figure 4 are also presented in a larger format in appendix D. The geology, pedology, DEMs, hydrology and slope map are is presented in Appendix B and the sub-indices of the 500m grid are included in Appendix C. The other sub-indices are presented in the Digital Appendix. The geodiversity patterns on the youngest and oldest island are briefly explained with the use of the input maps.
The highest geodiversity on the oldest island occurs in particular on areas with high levels of slope angles and elevation, containing entisols, oxisols, and many perennial rivers. Furthermore, very high geodiversity can be found on the Waimea Canyon Basalt. On the contrary, very low geodiversity values can be found on Koloa volcanics and alluvium.
Remarkable on the youngest islands, is that there is only a very small area in the north of Hawai’i Island showing very high geodiversity values. This area can be characterized to contain Polulu volcanics and Andisols. Some areas at the north coast of Hawai’i also contain high levels of geodiversity, characterized by an area of many perennial rivers. Comparable areas of this geodiversity class are characterized by Laupahoehoe volcanics, Hamakua volcanics and Tephra deposits. At sites with low geodiversity the soil are characterized by lava flows, histosols, Puna basalt or Kau basalt. The patterns of geodiversity are further discussed chapter 5, after selecting the most appropriate grid size and GDI formula.
4.1 Statistics
In this paragraph the results of the statistical analysis of the GDI maps are presented. Firstly, the standard deviations are presented to evaluate the three grid sizes. Afterwards, the best grid size is used to select the most appropriate GDI formula, by calculating correlation matrices and percentages of geodiversity of each island.
The standard deviation of each island and each GDI formula for the three different grid sizes are presented in table 6. In the table it is shown that the highest standard deviations are found for the 500m grid. Consequently, further analysis was only done for this grid size.
Table 6: The standard deviations of the seven Hawaiian Islands for three different grid sizes and three GDIs.
Island GDI1 GDI2 GDI3
500x500m kauai 1,076362 1,162128 1,100982 Oahu 1,107579 1,149421 1,102958 Molokai 0,869124 1,004803 0,898657 Lanai 1,005974 1,061743 0,952879 Maui 0,975188 1,11731 1,046632 Kahoolawe 0,776831 0,773383 0,740265 Big Island 0,742127 0,7327 0,806722 750x750m kauai 0,96915 0,897249 0,860881 Oahu 0,864559 0,971271 0,87203 Molokai 0,783779 0,870568 0,704018 Lanai 0,866491 0,948405 0,876559 Maui 0,934231 1,076284 0,999893 Kahoolawe 0,536235 0,41776 0,502846 Big Island 0,492602 0,795072 0,396853 1x1km kauai 0,876246 0,93111 0,833277 Oahu 0,897108 1,049358 0,957246 Molokai 0,942866 1,035266 0,892699 Lanai 1,09515 1,184785 1,04528 Maui 1,086901 1,058809 1,027724 Kahoolawe 0,555664 0,849892 0,619056 Big Island 0,592949 0,901164 0,829724
The outputs of the band collection statistics tool of the three different formulas of the 500m grid is presented in tables 7, 8 and 9. This takes into account the effect of each sub-index on the GDI of the whole archipelago
Table 7: The correlation matrix of GDI 1 (500m grid)
GDI1 Elevation diversity (std) geological diversity soil siversity slope diversity (std) hydrology diversity GDI1_500 1 Elevation diversity (std) 0,84 1 geological diversity 0,25 0,10 1 soil diversity 0,38 0,09 0,20 1 slope diversity (std) 0,75 0,56 0,16 0,21 1 hydrology diversity 0,56 0,28 0,08 0,24 0,41 1
Table 8: The correlation matrix of GDI2 (500m grid)
Table 9: The correlation matrix of GDI3 (500m grid)
GDI2 Elevation diversity (range) geological diversity soil siversity slope diversity (range) hydrology diversity GDI2_500 1 Elevation diversity (range) 0,84 1 geological diversity 0,25 0,13 1 soil diversity 0,34 0,10 0,20 1 slope diversity (range) 0,89 0,73 0,14 0,18 1 hydrology diversity 0,56 0,30 0,08 0,24 0,44 1 GDI3 Slope diversity (range) Slope diversity (std) elevation (range) elevation (std) geological diversity soil diversity hydrology diversity GDI3_500 1 Slope diversity (range) 0,84 1 Slope diversity (std) 0,75 0,84 1 elevation (range) 0,84 1 0,84 1 elevation (std) 0,8 0,67 0,56 0,67 1 geological diversity 0,24 0,14 0,16 0,14 0,1 1 soil diversity 0,35 0,18 0,21 0,18 0,1 0,2 1 hydrology diversity 0,59 0,44 0,41 0,44 0,28 0,1 0,24 1
The correlation matrices of the three GDIs do not show large differences. GDI3 shows the most
contrasting correlation values of the sub-indices on the GDI, in comparison to GDI1 and GDI2. GDI3 has also a correlation of 1 between elevation and slope range. This means that those two sub-indices are perfectly correlated, which may be the reason that it not appropriate to include both parameters for calculating the GDI since the effect of the sub-indices is exactly the same in this formula. Presented in in table 3 are the maximum geodiversity scores of GDI1 and GDI2 indicating a maximum diversity score of (18). On the contrary, GDI3 has a lower maximum diversity score (17,5), which indicates lower variety. Despite the high standard deviations of GDI2 (table 2) it can be suggested that GDI2 is not the most appropriate because of a very strong linear dependence of slope (0.89) and elevation range (0,84) on the geodiversity in comparison to the other indices as shown in table 8. The correlation matrices of GDI1 also show high correlations of elevation (0,84) and slope diversity (0,74), though not as strongly. Slope and elevation diversity of GDI1 do also show less correlation than in GDI2, what may result in higher variety since the effects of the sub-indices are less similar than for GDI2. That is why we will further analyse GDI1. Statistics of GDI1 are presented in table 10.
Table 10: The ages, means, percentages of all GDI3 classes of each island. Ages are obtained from Eckstut (2011).
The statistics of table 10, including age and percentages of the GDI1 classes per island, are also visualized in figure 5. The percentages of geodiversity hotspots per island are in a scatterplot (fig. 6).
Name island Age [Ma] Mean GDI3 very high GDI [%] High GDI [%] Moderate GDI [%] Low GDI [%] Very low GDI [%] Hawaii 0,6 3,112041 0,13 1,23 12,71 41,18 44,77 Kahoolawee 1,8 2,874278 0 1,79 21,15 44,22 32,84 Maui 1 2,449068 1,78 16,91 44,05 22,48 14,77 Lanai 1,6 2,562619 2,57 13,96 37,25 29,61 16,61 Molokai 2,1 2,690965 0,16 6,70 49,20 25,72 18,22 Oahu 4 1,917253 3,72 27,69 37,86 13,73 17,01 Kauai 5,3 1,708267 7,49 30,26 39,63 11,18 11,47 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Hawai'i [0,6 Ma] Kahoolawee [1.8 Ma] Maui'i [1 Ma] Lanai [1.6 Ma] Molokai [2.1 Ma] Oahu [4 Ma] Kaui'i [5.3 Ma]
Percentages of all GDI classes per island
Figure 5: Percentages of all GDI classes, from very low to very high, of each island in combination with its age.
Figure 6: Scatterplot of the percentages of very high geodiversity related to the age of the islands of the Hawaiian archipelago. The data points are labelled by the name of the islands.
Very high, high and moderate geodiversity are increasing, while low and very low are decreasing on the older islands (fig. 5). Figure 6 shows a trend in geodiversity from the younger to the older islands. This pattern is further discussed in chapter 5.
Hawaii
Maui
Kahoolaawee
Lanai
Molokai
Oahu
Kauai
0 5 10 15 20 25 0 1 2 3 4 5 6Ve
ry
high
G
D
I [%]
Age [Ma]
Chapter 5
Discussion
5.1 Methodology
For establishing the formulas in this research, the most used definition for assessing geodiversity of Gray (2004) was used as a basis. For other study areas, parameters can be removed or added to establish a geodiversity formula, suiting the need for the certain research. Selection of the best grid size depends on the aim of the study and the study area itself (Hengl, 2006).
The GDI is also dependent on the quality of the input data. The classification of the maps has a large influence on the GDI, since the variety of a single cell is calculated. The input data are expert-derived thematic maps created by geologists. As an area is classified in a more general way than the index value may be lower in comparison to when it is classified in a very detailed manner. The geology map contains some features with the same name, but they contain other properties, resulting in higher geodiversity scores in comparison to situations where the features with the same name are merged. An bias in the research is that the geology and soil map contain some water features. Besides, the soil map contains no available data at some areas. These features could not be replaced by a value of zero when calculating the GDI, because the zonal statistics tool did not had that option. When removing the invalid data completely, this resulted in an error.
The hydrology index can be further developed since this methodology deals with each river as only one feature. This could be improved by separating every river branch and giving each an individual weighting, to prevent generalization. The methodology can also be improved by adding or removing parameters, for example the slope aspect, as done by Hjort & Luoto (2010), taking variations in climate into account or by developing a geomorphology map of the Hawaiian Islands in order to include the genesis of the landforms as well. The application of different weightings for the parameters could also be considered.
5.2 Interpretation of the results
The results shows an increase in very high GDI from the younger islands, in early stages, towards the older islands, in later stages of the Wilson cycle, in north-western direction (fig. 6). Higher percentages of lower classes of geodiversity can be found on the younger islands (fig 5.). The hypotheses seems not to be true. Age and stage of a hotspot island in the Wilson cycle are closely related (Wilson, 1963). This means that we can suggest a relationship between the geodiversity and stage of island development.
It is considered that the factor time may be the main cause for the pattern in high geodiversity and not the topographic complexity of Whittaker et al. (2008). In particular, the sub-indices soil, hydrology, slope and elevation diversity used for the calculation of geodiversity are dependent on time. Appendix C shows an increase in variation of each sub-index in north-western direction. This pattern can be explained by several processes.
Firstly, the sub-index geological diversity is also dependent on time, because of the tectonic settings of the islands. For the young, active volcanic islands every lava flow will cause a unique geologic composition. This process will have less influence as islands are shifting further away from the hotspot.
Besides, a key factor in the geodiversity patterns may be the co-evolution of volcanic topography and hydrology. Hydrologic processes are key drivers of changes in volcanic landscapes. The evolution of the volcanic environment due to water flow affects in turn the hydrological processes (Jefferson et al.,
2014). The hydrology map presented in Appendix B is showing an increase in the amount of rivers as a
function of the total area from the young to the older islands, which is the reason for the higher hydrology diversity towards the older islands.
The low hydrology diversity on the young islands is in turn the reason for low variety in slope angles and elevations that can be found here, because the rivers on the younger islands are disrupted due to lava flows (Jefferson, 2014). Besides, lava flows on young islands do also disturb soil
development, since low geodiversity values are found here. More important is that young islands contain young soils resulting in low soil diversity.
However, since older islands have more stable conditions, the soils are older as well. When islands move further away from the volcanic active centre, degradation will play a more important role. Soil development, rock weathering and sediment deposition promote hydrologic changes as these processes generate more unconsolidated material (Jefferson et al., 2014). Hjort & Luoto (2010) also found higher geodiversity at sites where erosion played a major role. As more water is routed through shallow sub surfaces, more material becomes available for fluvial erosion. More erosion results in higher roughness of the surface (Jefferson et al., 2014) and thus higher variety of slope and elevation diversity. Topographic maps showing differences in drainage networks in combination with the geodiversity between the youngest (0.6 Ma) and oldest (5.3 Ma) Hawaiian islands are presented in figure 7.
Figure 7: Detailed geodiversity maps (GDI3, 500m grid) of the youngest (Hawai’i) and the oldest (Kaui’i) Hawaiian Island in comparison to the topographic maps (Jefferson et al., 2011).
5.3 Further research
This research may provide new perspectives on other hotspot island chains and the interpretation of their geodiversity in relation to the stage of island development. The geodiversity patterns may also play an important role in the biodiversity patterns. Further research could compare geodiversity hotspot with biodiversity hotspot on archipelagos.
However, the geodiversity is not examined throughout the whole Wilson cycle. Calculating the GDI for islands with stages before maximum growth is not possible for the Hawaiian archipelago, because the Hawaiian Island with the highest elevation is already the youngest island. Therefore it is suggested that for further research this method has to be applied on other hotspot island chains, for example on the Canary Islands, where it is possible to calculate GDI throughout the whole Wilson cycle. Further research could examine the differences in geodiversity between archipelagos (external), examine geodiversity at different tectonic settings or connect island geodiversity to seafloor geodiversity with the aim to explore geodiversity in a more global manner.
Chapter 6
Conclusion
A geodiversity index together with geodiversity maps for the Hawaiian archipelago was developed successfully. The aim of this study was to measure and compare the geodiversity of the Hawaiian Islands. Three grid sizes were tested (500m, 750m and 1km). For this study a grid size of 500 by 500 meter seems the most appropriate. Furthermore, three geodiversity formulas were examined, based on the most used definition of Gray (2004). There was no geomorphology map available, so the parameters slope diversity, elevation diversity and hydrology diversity were used to represent the variety of
topography in the area. GDI1, including only standard deviations, was suggested to be the most appropriate formula. The geodiversity index reveals an increasing geodiversity pattern in north-western direction towards the older islands. The factor time plays an important role in the understanding of this pattern.
It can be concluded that there is a relationship between the geodiversity of the Hawaiian islands and the stage of island development. The geodiversity is increasing throughout the Wilson cycle. The older the islands, the higher the geodiversity. However, conclusions of geodiversity in stages of the Wilson cycle before maximum growth cannot be done yet. It is therefore suggested that the method presented in this research has to be applied on other hotspot island chains. Besides, further research can be done to geodiversity at different tectonic situations and connecting the geodiversity of islands with the seafloor geodiversity in order to explore geodiversity on a more global scale.
Acknowledgements
The completion of this bachelor thesis could not have been possible without the assistance and good supervisory skills of Mr. Harry Seijmonsbergen. His contribution and encouragement is sincerely appreciated and gratefully acknowledged. Furthermore, I want to thank the U.S. Geological Survey (www.usgs.gov) and the University of Hawai’i (www.soest.hawaii.edu) for making many datasets of the Hawaiian Islands available, which was essential for carrying out this research. I greatly appreciate the University of Amsterdam for the opportunity to work at the GIS-studio, which is providing great opportunities for the computational support. Lastly, I want to thank Roy van Bragt for his excellent knowledge of the English language.
References
Ascione, A., Cinque, A., Miccadei, E., Villani, F., & Berti, C. (2008). The Plio-Quaternary uplift of the Apennine chain: new data from the analysis of topography and river valleys in Central
Italy. Geomorphology, 102(1), 105-118.
Burek, C. V., & Prosser, C. D. (2008). The history of geoconservation: an introduction. Geological Society, London, Special Publications, 300(1), 1-5.
Carracedo, J. C., Day, S., Guillou, H., Rodríguez Badiola, E., Canas, J. A., & Pérez Torrado, F. J. (1998). Hotspot volcanism close to a passive continental margin: the Canary Islands. Geological
Magazine, 135(5), 591-604.
Chambers, R. (2014). Development of a geodiversity index on a volcanic island: Tenerife, Canary Islands. (Master’s thesis). Retrieved on 13 april 2017, from
https://esc.fnwi.uva.nl/thesis/centraal/files/f608211680.pdf
Eckstut, M. E., McMahan, C. D., Crother, B. I., Ancheta, J. M., McLennan, D. A., & Brooks, D. R. (2011). PACT in practice: comparative historical biogeographic patterns and species–area relationships of the Greater Antillean and Hawaiian Island terrestrial biotas. Global Ecology and Biogeography, 20(4), 545-557.
Egger, A. E. & Carpi, A. (2010). “Scientific Controversy” Visionlearning Vol. POS-3 (1). Retrieved on 29 April 2017 from
http://www.visionlearning.com/en/library/Process-of-Science/49/Scientific-Controversy/181
Gray, M. (2004). Geodiversity: valuing and conserving abiotic nature. John Wiley & Sons. Greenberg, H. Hawaii 10m DEMs. U.S. Geological Survey, Seattle.
Hengl, T. (2006) Finding the right pixel size. Computers & Geosciences 32(9), 1283-1298. Hjort, J., & Luoto, M. (2010). Geodiversity of high-latitude landscapes in northern Finland. Geomorphology, 115(1), 109-116.
Jefferson, A. J., Ferrier, K. L., Perron, J. T., & Ramalho, R. (2014). Controls on the hydrological and topographic evolution of shield volcanoes and volcanic ocean islands. The Galapagos: A Natural Laboratory for the Earth Sciences, 204, 185.
Jenks, G. F. (1967). The Data Model Concept in Statistical Mapping. International Yearbook of Cartography 7: 186-190.
Kozlowski, S. (2004) Geodiversity. The concept and scope of geodiversity. Przeglad Geologiczny Selected Full Texts 52: 833-883.
Mergurian, C. & Okulewicz, S. (2007). Geology of Hawaii. Retrieved on 26 april 2017, from http://www.dukelabs.com/Publications/PubsPdf/HU0706_Geol280FHawaiiGuide.pdf NOAA, (2016). Climate of Hawai’i. Retrieved on 19 June 2017, from
http://www.prh.noaa.gov/hnl/pages/climate_summary.php
Pereira, D. I., Pereira, P., Brilha, J., & Santos, L. (2013). Geodiversity assessment of Paraná State (Brazil): an innovative approach. Environmental Management, 52(3), 541-552.
Press. F. & Siever, R. (1998). Understanding Earth, 2nd edition. Geological Magazine. 135(6), 819-842. Raharimahefa, T. (2012). Geoconservation and geodiversity for sustainable development in Madagascar. Madagascar Conservation & Development, 7(3), 126-134.
Ruszkiczay-Rüdiger, Z., Fodor, L., Horváth, E., & Telbisz, T. (2009). Discrimination of fluvial, eolian and neotectonic features in a low hilly landscape: A DEM-based morphotectonic analysis in the Central Pannonian Basin, Hungary. Geomorphology, 104(3), 203-217.
Serrano, E., & Ruiz-Flaño, P. (2007). Geodiversity: a theoretical and applied concept. Geographica Helvetica, 62(3), 140-147.
Serrano, E., Ruiz-Flaño, P., & Arroyo, P. (2009). Geodiversity assessment in a rural landscape: Tiermes-Caracena area (Soria, Spain). Memorie Descrittive Della Carta Geoligica d’Italia, 87, 173-180.
Sherrod, D. R., Sinton, J.M., Watkins, S. E. & Brunt, K. M. (2007). Geologic map of the State of Hawaii. U.S. Geological Survey, California.
Sparrow, G. W. A. (1965). Soils and Landforms. Proceedings of The South African Sugar Technologists' Association-March.
Snelling, A. (n.d.) De volcanic origin of the Hawaiian islands: rapidly creating a paradise. Retrieved on 7 may 2017, from http://www.origins.org.ua/page.php?id_story=1853
USDA (U.S. Department of Agriculture), (2015). Soil Survey Geographic (SSURGO) database for islands of Hawaii area. SSURGO, Texas.
USGS (2010). National Hydrography Dataset (NHD). National Geospatial Technical Operations Center
Whittaker, R. J., Triantis, K. A., & Ladle, R. J. (2008). A general dynamic theory of oceanic island biogeography. Journal of Biogeography, 35(6), 977-994.
Wilson, J. T. (1963). A possible origin of the Hawaiian Islands. Canadian Journal of Physics, 41(6), 863-870.
Zwoliński, Z., & Stachowiak, J. (2012). Geodiversity map of the Tatra National Park for geotourism. Quaestiones geographicae, 31(1), 99-107.
Appendices
Appendix A: The physiography of the Hawaiian Islands
Geological setting HawaiiFigure 1: The geological setting of the Hawaiian volcanic island chain. The Pacific plate is moving in north-western direction over a stationary hotspot (Snelling, n.d.).
The wilson cycle
Figure 2: Cross-section of the Canary islands showing age versus height. First the islands are increasing in height because of tectonic processes, after which the islands degrade to an almost flat surface due to erosion and subsidence. This is also referred to as the Wilson cycle (Carracedo et al. 1998).
